Natural streams typically have highly complex flow patterns. Velocity gradients, circulation zones, transverse flows, and other flow patterns are created in the presence of topographic features (e.g. exposed boulders, bars). How flow complexity influences a stream's ecological health and morphological stability, as well as how flow complexity responds to changes in hydrologic conditions, is poorly understood. One-dimensional (1-D) hydraulic models and two-dimensional (2-D) models that do not explicitly incorporate meso-scale topographic features are not capable of adequately reproducing the flow patterns found in channels having complex topography. Moreover, point measurements of depth and velocity, which are used to describe hydraulic conditions in habitat suitability studies, cannot be used to characterize spatially varying flow patterns of biological importance. A general methodology for incorporating meso-scale topography into 2-D hydraulic models is presented. The method provides a means of adequately reproducing spatial flows of interest to riverine researchers. The method is developed using 2-D model simulations of a reach of the North Fork of the Feather River in California. Specifically, the site is modeled with and without bathymetry data on exposed boulders found within the site. Results show that the incorporation of boulder topography and an adequately refined mesh are necessary for reproducing velocity gradients, transverse flows, and other spatial flows. These simulations are also used to develop and evaluate three spatial hydraulic metrics designed to distinguish between locations having uniform and non-uniform flow conditions. The first two metrics describe local variations in energy/velocity gradients, while the third metric provides a measure of the flow complexity occurring within an arbitrary area. The metrics based on principles of fluid mechanics (kinetic energy, vorticity, and circulation) can be computed in the field or with 2-D hydraulic model results. These three metrics, used in conjunction with detailed 2-D hydraulic model results, provide engineers, biologist, and water resource managers a set of tools with which to evaluate the importance of flow complexity within rivers. A conceptual model describing how such a tool can be used to help design channels being restored, better evaluate stream habitat, and evaluate how hydrologic changes in a watershed impact hydraulic conditions and concomitant habitat conditions is provided. / Ph. D.

Reservoir releases for generating power need to be reconciled with efforts to maintain healthy ecosystems in regulated rivers having irregular channel topography. Fluctuating, complex flow patterns near river obstructions such as boulders and large woody debris provide unique habitat for many aquatic organisms. Numerical modeling of the flow structures surrounding these obstructions is challenging, yet it represents an important tool for aquatic habitat assessment. Moreover, efforts for modeling the morphologically and biologically important transient flows, as well as quantifying their impacts on physical fish habitat during the unsteady-flow period remain rare. In this dissertation, the ability of two- (2-D) and three-dimensional (3-D) hydraulic models to reproduce the localized complex flow features at steady base and peak flows is examined first. The performance of the two hydraulic models is evaluated by comparing the numerical results with measurements of flow around a laboratory hemisphere and boulders located at a reach of the Smith River in Virginia. Close agreement between measured values and the velocity profiles predicted by the two models is obtained outside the wakes behind these obstructions. However, results suggest that in the vicinity of theses obstructions the 3-D model is better suited for reproducing the circulation flow behavior favored by many aquatic species over a broad range of flows. Further, time-dependent flow features affecting channel morphology and aquatic physical habitat are investigated using the numerical models for the same reach in the Smith River. Temporal variation measurements of water surface elevation and velocity profile obtained in the field during a reservoir release are in good agreement with the numerical results. A hypothetical "staggering" flow release scenario simulated by the 3-D model leads to reduced erosional area and longer refugia availability for juvenile brown trout during hydropeaking. Finally, an unsteadiness parameter β is proposed for determining whether an unsteady flow regime can be either modeled using a truly dynamic flow approach or a quasi-steady flow method. / Ph. D.